Estimate how many feature rows to inspect when measuring a defect, missing-value, or staleness proportion. Set an expected rate, desired margin of error, confidence level, and optional finite population to plan a defensible feature-store quality sample.
p is the expected proportion, e is the desired half-width, and N is the optional dataset size.
Use the main result together with the supporting bounds, counts, or capacity figures. The estimate is only as reliable as the input definitions, sampling process, and operating assumptions.
The formula assumes a random sample and independent observations. Clustered entities or repeated rows may require a larger design-adjusted sample.
At an expected 5% rate, ±1 percentage-point margin, and 95% confidence, the unadjusted formula requires 1,825 observations before any finite-population correction.
The calculator uses the normal approximation for a proportion, then applies a finite population correction when a dataset size is supplied.
Use 50% for the most conservative sample under this formula. A reliable pilot estimate can reduce the required sample.
Precision improves with the square root of sample size, so halving the margin generally requires about four times as many observations.
Finite population correction matters when the initial sample is a meaningful share of the entire dataset.
For a simple random-sample interpretation, yes. Stratified or clustered designs need weights or design-effect adjustments.
| Choice | Effect on sample |
|---|---|
| Higher confidence | Increases required sample |
| Smaller margin | Increases required sample substantially |
| Finite dataset | May reduce required sample |